First, the Federal government changed its industry classification system from the Standard Industrial Classification (SIC) to the North American Industry Classification System (NAICS). NAICS codes reflect recent technological changes, and also the growth and diversification of services.
The 2003 Report uses the NAICS codes. Previous reports used the 1987 SIC codes.
The second significant change is improved geographic information. In the 2003 Report, geocoding, a computerized way of finding the precise physical location of a postal address, was used to assign to each business the actual city, township, or unorganized territory in which the business was located. This results in a more accurate assignment of a business to a physical location.
The dollar amounts and counts for some central cities and cities with post offices are smaller than with the previous methodology. The change primarily affects city data, but county data are affected too.
Prior to the 2003 Report, city data were primarily based on the postal city of record for each business location. For some business with a postal city of Minneapolis or St. Paul, an effort was made to remove the data for suburban businesses from the statistics of the central cities and place the data in the statistics of the appropriate suburban city. This led to more accurate statistics for the cities involved, however no such adjustment was made for the rest of the state. So the data for some businesses in rural townships and small cities without a post office were placed in the statistics for larger central cities.
Data for some businesses with a centralized billing or distribution center will show up in the statistics for one central city even though the businesses are actually doing statewide or regional business from that single reporting location. The result is statistical data for that city is overstated in the sense that the good or service is not being delivered entirely in that city. So in estimating the amount of sales tax activity in a given location, the city statistics for some cities (those with centralized billing or distribution centers) would be too high and for other cities (those without) would be too low.
The growth of centralized billing and distribution centers for various goods and services has resulted in a more noticeable impact on city and county data.